There is no "treatment of choice" for patients with borderline disorder, a syndrome manifested by severe affective, schizotypal and impulsive-destructive symptoms. Despite a reported prevalence of up to 25% of admissions to some inpatient units, and a social and vocational prognosis akin to schizophrenia, there are few systematically controlled drug studies of borderline patients. To address this need, we are conducting the first double-blind placebo-controlled study of haloperidol (HAL) and amitriptyline (AMI) in criteria-defined borderline inpatients. Nearing completion of our study we find a statistically significant superiority for active medication over placebo (PLC) and for HAL over both AMI and PLC on a broad spectrum of measures representing affective and schizotypal symptom domains. HAL produces significant improvement on a composite measure of symptom severity compared to both PLC and AMI, while AMI differs little from PLC. The magnitude of overall improvement is modest though clinically and statistically significant following six weeks of acute treatment. Despite a high prevalence of depressive symptoms in our borderline sample, outcome with AMI differs little from PLC with small gains limited to areas of self-rated depressed mood. We find no relationship between response to AMI and a concurrent RDC diagnosis of major depression. AMI nonresponders show clinical and statistical evidence of serious behavioral toxicity. Pilot studies on the pattern of depression in our borderline patients indicate a prevalence of atypical depression and hysteroid dysphoria in 33-58% of patients, suggesting a role for MAOI antidepressants. Our follow-up studies document re-hospitalization within 16 weeks of 25% of study patients indicating a need for continuation therapy beyond acute inpatient treatment. To address these issues, we propose a) a 6 week double-blind, placebo-controlled acute treatment study comparing HAL to the MAOI antidepressant phenelzine in criteria defined borderline inpatients, followed by b) a 16 week outpatient continuation study of treatment responders. Together with the results of Pharmacotherapy I, this study will 1) define the first empirical guidelines for medication treatment of borderline patients, 2) define therapeutic and prophylactic efficacy in continuation treatment and 3) allow a pharmacologic behavioral dissection of borderline disorders into meaningful subtypes.